Project Summary/Abstract Blindness is devasting condition that impacts millions of people across the world. In most cases of adult-onset blindness there is damage or dysfunction of the eye, retina, or optic nerve, but the visual cortex is left intact. Direct electrical stimulation of visual cortex, even in blind patients, produces perception of distinct spots of light known as phosphenes. It has long been recognized that this could form the basis for a visual cortical prosthesis (VCP), a device which could greatly improve the quality of life for blind patients by restoring some visual function. In recent years, there has been renewed interest in development of VCPs due to technical advances in computing, wireless data and power transmission, and electrode arrays. While these advances have greatly improved the interface with the cerebral cortex, substantial research is still required to determine how to use that interface to communicate visual information using electrical stimulation. VCPs have typically used electrical stimulation delivered in a way that provides an unnatural input to the visual cortex. This includes using a sequence of electrical stimulation pulses delivered at an arbitrary fixed frequency and delivered in a manner that is unrelated to ongoing cortical activity. Electrical stimulation delivered in this unnatural fashion has been one of the key limitations in the development of effective VCPs. This proposal focuses on multiple ways to deliver electrical stimulation in a more naturalistic fashion. Aim 1 examines the importance of stochastic variability in the timing and amplitude of electrical stimulation pulses in generation of visual percepts. Aim 2 evaluates the importance of coordinating electrical stimulation relative to patterns of ongoing cortical activity. If naturalistic stimulation is more effective in generating visual percepts, this should be accompanied by more effective activation of visual pathways, and this will be explicitly tested in Aim 3. This will be done by combining electrical stimulation of visual cortex with functional imaging. Naturalistic electrical stimulation protocols could lower the current requirements for future VCPs, and therefore improve device safety and longevity. More generally, the results from these experiments may reveal general principals of how to effectively and efficiently encode information into the cerebral cortex in other brain computer interface applications. Furthermore, determining the most effective ways to input information into the brain using electrical stimulation could improve scientific understanding of normal mechanisms of cortical information processing and the relationship between cortical activity and perception.